A predictive model based on BRCA1/2, POLE, TP53, and MSH6 mutations for immunotherapy response in advanced endometrial cancer.

OBJECTIVE: To evaluate clinical, molecular, and immunological predictors of response to immunotherapy among patients with advanced endometrial cancer and to develop a combined biomarker model for predicting treatment outcomes. METHODS: This retrospective case-control study included 590 advanced endometrial cancer patients treated at the Affiliated Hospital of Hebei University of Engineering between December 2024 and May 2025. Eligible women underwent total hysterectomy, pelvic lymph node dissection, and received immune checkpoint inhibitors alongside standard chemotherapy. Patients were stratified into good and poor response groups based on 1-year post-treatment prognosis and response evaluation criteria in solid tumors. Baseline blood biomarkers, gene mutation status (breast cancer gene [BRCA] 1, BRCA2, DNA polymerase epsilon, tumor protein p53 [TP53], mutS homolog 6), and immunophenoscore (IPS) were assessed. Logistic regression and receiver operating characteristic (ROC) analyses were performed. A random forest model was constructed for combined biomarker prediction. RESULTS: No significant differences in baseline demographic or clinical characteristics were found between response groups. Good responders had significantly lower baseline levels of C-reactive protein (CRP), interleukin-6 (IL-6), tumor necrosis factor alpha (TNF-α), neutrophil-lymphocyte ratio (NLR), cancer antigen 125 (CA125), and IPS, and higher frequencies of gene mutations. Multivariate regression identified elevated CRP, IL-6, TNF-α, NLR, CA125, and IPS as independent predictors of poor response; BRCA2 and TP53 mutations were independently associated with favorable outcomes. The combined biomarker model achieved an area under the ROC curve of 0.812, demonstrating strong predictive accuracy. CONCLUSION: Inflammatory and tumor biomarkers, IPS, and specific gene mutations are independently associated with immunotherapy response in advanced endometrial cancer. A combined biomarker model may enhance the prediction of treatment outcomes and guide individualized therapy.